HPC YEARBOOK 2021/22


Increasing hardware diversity


Kalyan Kumaran, Dr Sergi Girona, and professor Simon McIntosh- Smith discuss hardware diversity and its impact on scientists


Can you give me your thoughts on hardware diversity and what this means for scientists and researchers?


Kalyan Kumaran, senior computer scientist and the director of technology at Argonne’s Leadership Computing Facility (ALCF): Our craft at ALCF is to find a platform that accelerates scientific insight. We want faster science, leading to higher productivity. But in doing that, we don’t want to restrict ourselves to a few of the top-notch vendors who provide HPC solutions. We want to explore a wide variety of vendors and a diverse set of platforms.


That creates “


more uncertainty, and an awful lot more analysis that has to be done by almost everyone


“ We ensure that we get the best bang


for the buck or the best computing performance for each watt. We want to select the technology that can run our particular workloads, and we use several factors to do that. The maturity of the software, how easy it is to use, the performance of the hardware, how easy is it to access the systems and how well it fits into a facility. However, the most important thing


for us when supporting our users or supporting our scientists is to make sure they are using the latest open- source programming models or open- source performance tools. We don’t want scientists to be performance engineers that spend


all their time moving their science applications from one platform to another. We want them to be focused on science. We want them to uncover new algorithms, to do things faster and solve bigger problems. But to do that, we make sure that


they are using open programming models. And we make sure that these programming models are developed by being part of the standards committees that develop them. We make sure these programming models are supported, even if we change our vendor platforms. And, by supported, I mean it’s not just functional but that they perform very well.


>


4


www.scientific-computing.com


Andrey Suslov/shutterstock


Page 1  |  Page 2  |  Page 3  |  Page 4  |  Page 5  |  Page 6  |  Page 7  |  Page 8  |  Page 9  |  Page 10  |  Page 11  |  Page 12  |  Page 13  |  Page 14  |  Page 15  |  Page 16  |  Page 17  |  Page 18  |  Page 19  |  Page 20  |  Page 21  |  Page 22  |  Page 23  |  Page 24  |  Page 25  |  Page 26  |  Page 27  |  Page 28  |  Page 29  |  Page 30  |  Page 31  |  Page 32  |  Page 33  |  Page 34  |  Page 35  |  Page 36  |  Page 37  |  Page 38  |  Page 39  |  Page 40  |  Page 41  |  Page 42